ref: .fit hook clean up (#3198)

* eval loop clean up

* eval loop clean up

* eval loop clean up

* eval loop clean up

pytorch_lightning/trainer/evaluation_loop.py

@@ -245,14 +245,11 @@ class TrainerEvaluationLoopMixin(ABC):
         # hook
         self.evaluation_loop.on_evaluation_start()
 
-        # ------------------------------
-        # ------------------------------
-        # ------------------------------
         # set up the eval loop
         self.evaluation_loop.setup(model, max_batches, dataloaders)
 
         # hook
-        # TODO: needs to move inside the loop but breaks early stopping
+        # TODO: should this be insider the dataloader loop?
         self.evaluation_loop.on_evaluation_epoch_start()
 
         # run validation/testing
@@ -300,10 +297,6 @@ class TrainerEvaluationLoopMixin(ABC):
         # hook
         self.evaluation_loop.on_evaluation_epoch_end()
 
-        # ------------------------------
-        # ------------------------------
-        # ------------------------------
-
         # log the final eval loop metrics
         eval_loop_results = self.__log_evaluation_epoch_metrics(eval_results, test_mode)
 

pytorch_lightning/trainer/trainer.py

@@ -995,10 +995,8 @@ class Trainer(
         # check that model is configured correctly
         self.config_validator.verify_loop_configurations(model)
 
-        # callbacks
-        self.on_fit_start(model)
-        if self.is_function_implemented('on_fit_start', model):
-            model.on_fit_start()
+        # hook
+        self.call_hook('on_fit_start', model)
 
         # on multi-gpu jobs we only want to manipulate (download, etc) on node_rank=0, local_rank=0
         # or in the case where each node needs to do its own manipulation in which case just local_rank=0
@@ -1095,12 +1093,10 @@ class Trainer(
             self.accelerator_backend.setup(model)
             results = self.accelerator_backend.train(model)
 
-        # on fit end callback
-        self.on_fit_end()
-        if self.is_function_implemented('on_fit_end'):
-            model.on_fit_end()
+        # hook
+        self.call_hook('on_fit_end')
 
-        # teardown callback
+        # hook
         self.teardown('fit')
         if self.is_function_implemented('teardown'):
             model.teardown('fit')